視覺化的語言心法—資料、模型、與溝通

How to use this slides

• Press Esc to navigation mode
• Press pageDown, Right or Down to go to next slide
• Presss PageUP, Left or Up to go to previous slide

一個例子學會畫圖：mpg 🚗油耗資料

• mpg dataset: Fuel economy data from 1999 and 2008 for 38 popular models of car.

Aesthetic Mappings

ggplot(data = <DATA>) + # Data
  geom_<xxx>(
     mapping = aes(<MAPPINGS>), ##  <= Aesthetic mappings
     stat = <STAT>,
     position = <POSITION>
  ) +
  scale_<xxx>() + coord_<xxx>() + facet_<xxx>()
  theme_()

Aesthetics 基本用法

• x = displ
• y = hwy
• color = class

ggplot(data = mpg) +
  geom_point(mapping = aes(x = displ, y = hwy, color = class), size = 3)

Aesthetics 基本用法

• x = displ
• y = hwy
• shape = class

ggplot(data = mpg) +
  geom_point(mapping = aes(x = displ, y = hwy, shape = class))

Layers 圖層觀念

• 兩層圖層

ggplot(data = mpg) +
  geom_point(mapping = aes(x = displ, y = hwy)) +
  geom_smooth(mapping= aes(x = displ, y = hwy))

Too many variables!!!!

• 多了第3個變數，所以更難理解了

Facets: Small-Multiples

• Facets 是很重要的一個呈現方式，一定要學起來
• 為什麼要用 Facets？
  1. 當同一個座標平面塞入太多變數，會造成大腦無法負荷
  2. 分拆資訊，讓大腦協助腦補更有效率

ggplot(data = mpg) +
  geom_bar(mapping = aes(x = class)) +
  facet_wrap( ~ manufacturer, ncol = 3)

(back to slides)

Take a look at our data

library(gapminder)
gapminder

#> # A tibble: 1,704 x 6
#>    country     continent  year lifeExp      pop gdpPercap
#>    <fct>       <fct>     <int>   <dbl>    <int>     <dbl>
#>  1 Afghanistan Asia       1952    28.8  8425333      779.
#>  2 Afghanistan Asia       1957    30.3  9240934      821.
#>  3 Afghanistan Asia       1962    32.0 10267083      853.
#>  4 Afghanistan Asia       1967    34.0 11537966      836.
#>  5 Afghanistan Asia       1972    36.1 13079460      740.
#>  6 Afghanistan Asia       1977    38.4 14880372      786.
#>  7 Afghanistan Asia       1982    39.9 12881816      978.
#>  8 Afghanistan Asia       1987    40.8 13867957      852.
#>  9 Afghanistan Asia       1992    41.7 16317921      649.
#> 10 Afghanistan Asia       1997    41.8 22227415      635.
#> # ... with 1,694 more rows

Our ggplot

R packages

1. Nested data (tidyr)
2. Functional programming (purrr)
3. Models → tidy data (broom)

Split our data into data.frames by group

Split our data

library(dplyr)
library(tidyr)

gapminder <- gapminder %>% mutate(year1950 = year - 1950)
by_country <- gapminder %>% 
  group_by(continent, country) %>% 
  nest

by_country

#> # A tibble: 142 x 3
#>    continent country     data             
#>    <fct>     <fct>       <list>           
#>  1 Asia      Afghanistan <tibble [12 × 5]>
#>  2 Europe    Albania     <tibble [12 × 5]>
#>  3 Africa    Algeria     <tibble [12 × 5]>
#>  4 Africa    Angola      <tibble [12 × 5]>
#>  5 Americas  Argentina   <tibble [12 × 5]>
#>  6 Oceania   Australia   <tibble [12 × 5]>
#>  7 Europe    Austria     <tibble [12 × 5]>
#>  8 Asia      Bahrain     <tibble [12 × 5]>
#>  9 Asia      Bangladesh  <tibble [12 × 5]>
#> 10 Europe    Belgium     <tibble [12 × 5]>
#> # ... with 132 more rows

Fit a model within each country

lm(lifeExp ~ year, data = Afghanistan)
lm(lifeExp ~ year, data = Afghanistan)
...

We can use purrr::map() to fit each model

library(purrr)

country_model <- function(df) {
  lm(lifeExp ~ year1950, data = df)
}

models <- by_country %>%
  mutate(mod_lm = map(data, country_model))

models

#> # A tibble: 142 x 4
#>    continent country     data              mod_lm  
#>    <fct>     <fct>       <list>            <list>  
#>  1 Asia      Afghanistan <tibble [12 × 5]> <S3: lm>
#>  2 Europe    Albania     <tibble [12 × 5]> <S3: lm>
#>  3 Africa    Algeria     <tibble [12 × 5]> <S3: lm>
#>  4 Africa    Angola      <tibble [12 × 5]> <S3: lm>
#>  5 Americas  Argentina   <tibble [12 × 5]> <S3: lm>
#>  6 Oceania   Australia   <tibble [12 × 5]> <S3: lm>
#>  7 Europe    Austria     <tibble [12 × 5]> <S3: lm>
#>  8 Asia      Bahrain     <tibble [12 × 5]> <S3: lm>
#>  9 Asia      Bangladesh  <tibble [12 × 5]> <S3: lm>
#> 10 Europe    Belgium     <tibble [12 × 5]> <S3: lm>
#> # ... with 132 more rows

What can we extract from each model?

Extract info from model

models <- models %>% 
  mutate(
    glance = mod_lm %>% map(broom::glance),
    tidy = mod_lm %>% map(broom::tidy),
    augment = mod_lm %>% map(broom::augment),
    rsq = glance %>% map_dbl("r.squared")
  )
models

#> # A tibble: 142 x 8
#>    continent country     data     mod_lm  glance   tidy    augment     rsq
#>    <fct>     <fct>       <list>   <list>  <list>   <list>  <list>    <dbl>
#>  1 Asia      Afghanistan <tibble… <S3: l… <data.f… <data.… <data.fr… 0.948
#>  2 Europe    Albania     <tibble… <S3: l… <data.f… <data.… <data.fr… 0.911
#>  3 Africa    Algeria     <tibble… <S3: l… <data.f… <data.… <data.fr… 0.985
#>  4 Africa    Angola      <tibble… <S3: l… <data.f… <data.… <data.fr… 0.888
#>  5 Americas  Argentina   <tibble… <S3: l… <data.f… <data.… <data.fr… 0.996
#>  6 Oceania   Australia   <tibble… <S3: l… <data.f… <data.… <data.fr… 0.980
#>  7 Europe    Austria     <tibble… <S3: l… <data.f… <data.… <data.fr… 0.992
#>  8 Asia      Bahrain     <tibble… <S3: l… <data.f… <data.… <data.fr… 0.967
#>  9 Asia      Bangladesh  <tibble… <S3: l… <data.f… <data.… <data.fr… 0.989
#> 10 Europe    Belgium     <tibble… <S3: l… <data.f… <data.… <data.fr… 0.995
#> # ... with 132 more rows

See the R^2

models %>% 
  ggplot(aes(rsq, reorder(country, rsq))) +  # use factor levels
  geom_point(aes(colour = continent)) +
  theme(axis.text=element_text(size=rel(0.7)))

Unnest to a regular data frame

models %>% unnest(data)

#> # A tibble: 1,704 x 8
#>    continent country       rsq  year lifeExp      pop gdpPercap year1950
#>    <fct>     <fct>       <dbl> <int>   <dbl>    <int>     <dbl>    <dbl>
#>  1 Asia      Afghanistan 0.948  1952    28.8  8425333      779.       2.
#>  2 Asia      Afghanistan 0.948  1957    30.3  9240934      821.       7.
#>  3 Asia      Afghanistan 0.948  1962    32.0 10267083      853.      12.
#>  4 Asia      Afghanistan 0.948  1967    34.0 11537966      836.      17.
#>  5 Asia      Afghanistan 0.948  1972    36.1 13079460      740.      22.
#>  6 Asia      Afghanistan 0.948  1977    38.4 14880372      786.      27.
#>  7 Asia      Afghanistan 0.948  1982    39.9 12881816      978.      32.
#>  8 Asia      Afghanistan 0.948  1987    40.8 13867957      852.      37.
#>  9 Asia      Afghanistan 0.948  1992    41.7 16317921      649.      42.
#> 10 Asia      Afghanistan 0.948  1997    41.8 22227415      635.      47.
#> # ... with 1,694 more rows

Unnest to a regular data frame

models %>% unnest(glance, .drop = TRUE)

#> # A tibble: 142 x 14
#>    continent country       rsq r.squared adj.r.squared sigma statistic
#>    <fct>     <fct>       <dbl>     <dbl>         <dbl> <dbl>     <dbl>
#>  1 Asia      Afghanistan 0.948     0.948         0.942 1.22      181. 
#>  2 Europe    Albania     0.911     0.911         0.902 1.98      102. 
#>  3 Africa    Algeria     0.985     0.985         0.984 1.32      662. 
#>  4 Africa    Angola      0.888     0.888         0.877 1.41       79.1
#>  5 Americas  Argentina   0.996     0.996         0.995 0.292    2246. 
#>  6 Oceania   Australia   0.980     0.980         0.978 0.621     481. 
#>  7 Europe    Austria     0.992     0.992         0.991 0.407    1261. 
#>  8 Asia      Bahrain     0.967     0.967         0.963 1.64      291. 
#>  9 Asia      Bangladesh  0.989     0.989         0.988 0.977     930. 
#> 10 Europe    Belgium     0.995     0.995         0.994 0.293    1822. 
#> # ... with 132 more rows, and 7 more variables: p.value <dbl>, df <int>,
#> #   logLik <dbl>, AIC <dbl>, BIC <dbl>, deviance <dbl>, df.residual <int>

Unnest to a regular data frame

models %>% unnest(tidy, .drop = TRUE)

#> # A tibble: 284 x 8
#>    continent country       rsq term  estimate std.error statistic  p.value
#>    <fct>     <fct>       <dbl> <chr>    <dbl>     <dbl>     <dbl>    <dbl>
#>  1 Asia      Afghanistan 0.948 (Int…   29.4     0.699       42.0  1.40e-12
#>  2 Asia      Afghanistan 0.948 year…    0.275   0.0205      13.5  9.84e- 8
#>  3 Europe    Albania     0.911 (Int…   58.6     1.13        51.7  1.79e-13
#>  4 Europe    Albania     0.911 year…    0.335   0.0332      10.1  1.46e- 6
#>  5 Africa    Algeria     0.985 (Int…   42.2     0.756       55.8  8.22e-14
#>  6 Africa    Algeria     0.985 year…    0.569   0.0221      25.7  1.81e-10
#>  7 Africa    Angola      0.888 (Int…   31.7     0.804       39.4  2.63e-12
#>  8 Africa    Angola      0.888 year…    0.209   0.0235       8.90 4.59e- 6
#>  9 Americas  Argentina   0.996 (Int…   62.2     0.167      372.   4.80e-22
#> 10 Americas  Argentina   0.996 year…    0.232   0.00489     47.4  4.22e-13
#> # ... with 274 more rows

Plot the models

models %>% 
  unnest(tidy) %>% 
  select(continent, country, term, estimate, rsq) %>% 
  spread(key = term, value = estimate) %>% 
  ggplot(aes(`(Intercept)`, year1950)) +
  geom_point(aes(colour = continent, size = rsq)) +
  geom_smooth(se = FALSE)

Plot the models: augmented

models %>% 
  unnest(augment) %>% 
  ggplot(aes(year1950, .resid)) +
  geom_line(aes(group = country), alpha = 0.3) +
  geom_smooth(se = FALSE) +
  geom_hline(yintercept = 0, colour = "red", alpha = 0.7) +
  facet_wrap(~continent)

Quick recap for model viz

Functional programming is very powerful

1. tidyr: 把物件 (例如 lm) 用 list 存在 columns 裡面
2. purrr: Functional programming
3. broom: Models → tidy data

Reference

• Hadley Wickham: Managing many models with R
• PLOTCON 2016: Hadley Wickham, New open viz in R
• Hadley Wickham “Data Science with R”
• Cookbook for R